← Back to community benchmarks

gemma-4-26b-a4b-it

M2 Max (38c) · 64 GB · 8bit · 2026-05-31
Performance
195k
tokens
307.1
PP tok/s
15.9
TG tok/s
651219
TTFT (ms)
38.6
Peak mem (GB)
Hardware
Chip M2 Max (38c)
Memory 64 GB
GPU Cores 38
Software
oMLX v0.3.12
macOS macOS 26.5
Context 200,000
Performance by Context Length
Context PP tok/s TG tok/s Peak Mem
1k 565.6 54.0 25.8 GB view
4k 616.9 51.9 26.4 GB view
8k 622.1 49.0 26.6 GB view
16k 606.2 44.5 27.1 GB view
32k 564.1 38.8 28.0 GB view
64k 497.2 30.5 30.0 GB view
128k 386.7 21.2 34.2 GB view
195k 307.1 15.9 38.6 GB current
Batching Results
Batch Size TG tok/s Speedup
54.0 1.00×
79.8 1.48×
103.7 1.92×